Suicide prediction models: a critical review of recent research with recommendations for the way forward

The dexamethasone suppression test as a biomarker for suicidal behavior: A systematic review and meta-analysis

BACKGROUND

The dexamethasone suppression test (DST), which measures HPA-axis functioning, is a potential biomarker for suicidal behavior. The current study aimed (a) to synthesize available knowledge on the association between DST non-suppression and suicidal behavior, and (b) to study potential moderators.

METHODS

A total of 4236 studies were screened, 43 were included. Suicide attempts and suicide completion were studied separately. The meta-analysis included 37 effect sizes for suicide attempts (n = 3733) and 11 effect sizes for suicide completion (n = 1626).

RESULTS

DST non-suppression was associated with completed suicide (odds ratio (OR) = 2.10, (95 % CI [1.37, 3.23]). For suicide attempts, we found no evidence that DST status was associated in the overall meta-analysis including all patient samples. However, moderator analysis indicated that the DST status was associated with suicide attempts in patient samples that included psychopathology other than just mood disorders, such as psychotic, substance use and personality disorders (OR = 2.34, 95 % CI [1.39-3.93], k = 11).

LIMITATIONS

The potential influence of publication bias and exclusion of some relevant published studies (since effect sizes could not be calculated, authors could not supply data or authors could not be reached) are limitations. Furthermore, missing moderator data decreased our ability to explain heterogeneity between studies.

CONCLUSIONS

The results of this meta-analysis support the hypothesis that DST non-suppression is predictive of suicidal behavior. More research is needed to investigate optimal cut-off values, confounding factors and the potential usefulness of the DST in clinical practice in terms of personalized medicine.

Clinical Risk Management in Mental Health Services: 10 Principles for Best Practice

Risk assessment and management are a fundamental part of clinical practice globally within mental health services. In the United Kingdom (UK), the evidence to support the effectiveness of structured risk assessment and management remains limited, although the perception remains that structured management frameworks are effective in reducing risk in mental health care. Despite the importance of risk management within mental health services, the most recent UK wide guidance was published in 2009, while international guidance for the assessing and management of service user risks also appears sparse. This perspective paper reports on a consultation and co-production project to provide up-to-date best practice principles in clinical risk management to enhance the consistency, quality and safety of mental health practice in the UK mental health services, and for mental health services in other English speaking countries. A three-stage approach was used including literature review, referral to mental health experts for review and final evaluation and sign off by users of mental health services as experts by experience. Ten principles for best practice were confirmed as a benchmark for practice and are offered as a benchmark to improve the quality and safety of mental health practice.

Detecting suicide risk among U.S. servicemembers and veterans: a deep learning approach using social media data

BACKGROUND

Military Servicemembers and Veterans are at elevated risk for suicide, but rarely self-identify to their leaders or clinicians regarding their experience of suicidal thoughts. We developed an algorithm to identify posts containing suicide-related content on a military-specific social media platform.

METHODS

Publicly-shared social media posts (n = 8449) from a military-specific social media platform were reviewed and labeled by our team for the presence/absence of suicidal thoughts and behaviors and used to train several machine learning models to identify such posts.

RESULTS

The best performing model was a deep learning (RoBERTa) model that incorporated post text and metadata and detected the presence of suicidal posts with relatively high sensitivity (0.85), specificity (0.96), precision (0.64), F1 score (0.73), and an area under the precision-recall curve of 0.84. Compared to non-suicidal posts, suicidal posts were more likely to contain explicit mentions of suicide, descriptions of risk factors (e.g. depression, PTSD) and help-seeking, and first-person singular pronouns.

CONCLUSIONS

Our results demonstrate the feasibility and potential promise of using social media posts to identify at-risk Servicemembers and Veterans. Future work will use this approach to deliver targeted interventions to social media users at risk for suicide.

Using natural language processing to evaluate temporal patterns in suicide risk variation among high-risk Veterans

Measuring suicide risk fluctuation remains difficult, especially for high-suicide risk patients. Our study addressed this issue by leveraging Dynamic Topic Modeling, a natural language processing method that evaluates topic changes over time, to analyze high-suicide risk Veterans Affairs patients' unstructured electronic health records. Our sample included all high-risk patients that died (cases) or did not (controls) by suicide in 2017 and 2018. Cases and controls shared the same risk, location, and treatment intervals and received nine months of mental health care during the year before the relevant end date. Each case was matched with five controls. We analyzed case records from diagnosis until death and control records from diagnosis until matched case's death date. Our final sample included 218 cases and 943 controls. We analyzed the corpus using a Python-based Dynamic Topic Modeling algorithm. We identified five distinct topics, "Medication," "Intervention," "Treatment Goals," "Suicide," and "Treatment Focus." We observed divergent change patterns over time, with pathology-focused care increasing for cases and supportive care increasing for controls. The case topics tended to fluctuate more than the control topics, suggesting the importance of monitoring lability. Our study provides a method for monitoring risk fluctuation and strengthens the groundwork for time-sensitive risk measurement.

Incidence rate and risk factors for suicide in patients with breast cancer in the USA: A surveillance, epidemiology, and end results analysis (SEER)

PURPOSE

To investigate suicide mortality and the related factors among female breast cancer patients in the United States.

METHODS

The SEER database was used to identify 716,422 patients diagnosed with breast cancer between 2010 and 2018 to calculate a standardized mortality rate (SMR). An analysis of risk factors for suicide death was conducted using the univariate and multivariate Cox proportional risk model. An estimation of suicide probability was performed through a nomogram model.

RESULTS

Compared with the expected suicide cases (n = 155) in the general population of the United States at the corresponding period (a suicide death rate of 5.71 per 100,000 person-years), the suicide rate among 716,422 breast cancer patients was followed during 2010-2018 and showed a relatively higher rate of 9.02 per 100,000 person-years. The SMR was 1.58 (95%CI: 1.39-1.79). White and other races were nine and seven times more likely to complete suicide than Black race, respectively (aHR = 9.013, 95%CI: 3.335-24.36, P < 0.001; aHR = 7.129, 95%CI: 2.317-21.931, P = 0.001); unmarried or single patients were at higher risk than married patients (aHR = 1.693, 95%CI: 1.206-2.377, P = 0.002). Patients receiving radiotherapy (aHR = 0.731, 95%CI: 0.545-0.980, P = 0.036) were less likely to complete suicide than those who did not.

CONCLUSION

Female breast cancer patients in the United States have a higher suicide rate than the general public, and the risk factors consist of non-black ethnicity, being single or unmarried, and not being treated with radiotherapy. As a result of this study, clinicians may be able to identify female breast cancer patients who are at high risk of suicide, thus providing appropriate psychological support at the early stage.

Validation of a Multivariable Model to Predict Suicide Attempt in a Mental Health Intake Sample

Importance

Given that suicide rates have been increasing over the past decade and the demand for mental health care is at an all-time high, targeted prevention efforts are needed to identify individuals seeking to initiate mental health outpatient services who are at high risk for suicide. Suicide prediction models have been developed using outpatient mental health encounters, but their performance among intake appointments has not been directly examined.

Objective

To assess the performance of a predictive model of suicide attempts among individuals seeking to initiate an episode of outpatient mental health care.

Design, Setting, and Participants

This prognostic study tested the performance of a previously developed machine learning model designed to predict suicide attempts within 90 days of any mental health outpatient visit. All mental health intake appointments scheduled between January 1, 2012, and April 1, 2022, at Kaiser Permanente Northern California, a large integrated health care delivery system serving over 4.5 million patients, were included. Data were extracted and analyzed from August 9, 2022, to July 31, 2023.

Main Outcome and Measures

Suicide attempts (including completed suicides) within 90 days of the appointment, determined by diagnostic codes and government databases. All predictors were extracted from electronic health records.

Results

The study included 1 623 232 scheduled appointments from 835 616 unique patients. There were 2800 scheduled appointments (0.17%) followed by a suicide attempt within 90 days. The mean (SD) age across appointments was 39.7 (15.8) years, and most appointments were for women (1 103 184 [68.0%]). The model had an area under the receiver operating characteristic curve of 0.77 (95% CI, 0.76-0.78), an area under the precision-recall curve of 0.02 (95% CI, 0.02-0.02), an expected calibration error of 0.0012 (95% CI, 0.0011-0.0013), and sensitivities of 37.2% (95% CI, 35.5%-38.9%) and 18.8% (95% CI, 17.3%-20.2%) at specificities of 95% and 99%, respectively. The 10% of appointments at the highest risk level accounted for 48.8% (95% CI, 47.0%-50.6%) of the appointments followed by a suicide attempt.

Conclusions and Relevance

In this prognostic study involving mental health intakes, a previously developed machine learning model of suicide attempts showed good overall classification performance. Implementation research is needed to determine appropriate thresholds and interventions for applying the model in an intake setting to target high-risk cases in a manner that is acceptable to patients and clinicians.

Uncommon Protein-Coding Variants Associated With Suicide Attempt in a Diverse Sample of U.S. Army Soldiers

BACKGROUND

Suicide is a societal and public health concern of global scale. Identifying genetic risk factors for suicide attempt can characterize underlying biology and enable early interventions to prevent deaths. Recent studies have described common genetic variants for suicide-related behaviors. Here, we advance this search for genetic risk by analyzing the association between suicide attempt and uncommon variation exome-wide in a large, ancestrally diverse sample.

METHODS

We sequenced whole genomes of 13,584 soldiers from the Army STARRS (Army Study to Assess Risk and Resilience in Servicemembers), including 979 individuals with a history of suicide attempt. Uncommon, nonsilent protein-coding variants were analyzed exome-wide for association with suicide attempt using gene-collapsed and single-variant analyses.

RESULTS

We identified 19 genes with variants enriched in individuals with history of suicide attempt, either through gene-collapsed or single-variant analysis (Bonferroni padjusted < .05). These genes were CIB2, MLF1, HERC1, YWHAE, RCN2, VWA5B1, ATAD3A, NACA, EP400, ZNF585A, LYST, RC3H2, PSD3, STARD9, SGMS1, ACTR6, RGS7BP, DIRAS2, and KRTAP10-1. Most genes had variants across multiple genomic ancestry groups. Seventeen of these genes were expressed in healthy brain tissue, with 9 genes expressed at the highest levels in the brain versus other tissues. Brains from individuals deceased from suicide aberrantly expressed RGS7BP (padjusted = .035) in addition to nominally significant genes including YWHAE and ACTR6, all of which have reported associations with other mental disorders.

CONCLUSIONS

These results advance the molecular characterization of suicide attempt behavior and support the utility of whole-genome sequencing for complementing the findings of genome-wide association studies in suicide research.

Predicting the Population Risk of Suicide Using Routinely Collected Health Administrative Data in Quebec, Canada: Model-Based Synthetic Estimation Study

BACKGROUND

Suicide is a significant public health issue. Many risk prediction tools have been developed to estimate an individual's risk of suicide. Risk prediction models can go beyond individual risk assessment; one important application of risk prediction models is population health planning. Suicide is a result of the interaction among the risk and protective factors at the individual, health care system, and community levels. Thus, policy and decision makers can play an important role in suicide prevention. However, few prediction models for the population risk of suicide have been developed.

OBJECTIVE

This study aims to develop and validate prediction models for the population risk of suicide using health administrative data, considering individual-, health system-, and community-level predictors.

METHODS

We used a case-control study design to develop sex-specific risk prediction models for suicide, using the health administrative data in Quebec, Canada. The training data included all suicide cases (n=8899) that occurred from January 1, 2002, to December 31, 2010. The control group was a 1% random sample of living individuals in each year between January 1, 2002, and December 31, 2010 (n=645,590). Logistic regression was used to develop the prediction models based on individual-, health care system-, and community-level predictors. The developed model was converted into synthetic estimation models, which concerted the individual-level predictors into community-level predictors. The synthetic estimation models were directly applied to the validation data from January 1, 2011, to December 31, 2019. We assessed the performance of the synthetic estimation models with four indicators: the agreement between predicted and observed proportions of suicide, mean average error, root mean square error, and the proportion of correctly identified high-risk regions.

RESULTS

The sex-specific models based on individual data had good discrimination (male model: C=0.79; female model: C=0.85) and calibration (Brier score for male model 0.01; Brier score for female model 0.005). With the regression-based synthetic models applied in the validation data, the absolute differences between the synthetic risk estimates and observed suicide risk ranged from 0% to 0.001%. The root mean square errors were under 0.2. The synthetic estimation model for males correctly predicted 4 of 5 high-risk regions in 8 years, and the model for females correctly predicted 4 of 5 high-risk regions in 5 years.

CONCLUSIONS

Using linked health administrative databases, this study demonstrated the feasibility and the validity of developing prediction models for the population risk of suicide, incorporating individual-, health system-, and community-level variables. Synthetic estimation models built on routinely collected health administrative data can accurately predict the population risk of suicide. This effort can be enhanced by timely access to other critical information at the population level.

Determining the MMPI-3 SUI scale's cross-sectional and prospective utility in suicide risk assessment

OBJECTIVE

In-depth suicide risk assessments are particularly important to long-term suicide prevention. Broadband measures of psychopathology, such as the Minnesota Multiphasic Personality Inventory (MMPI) instruments, assess suicide risk factors and various mental health comorbidities. With the recent release of the MMPI-3, the Suicidal/Death Ideation (SUI) scale underwent revisions to improve its construct validity and detection of suicide risk factors. Thus, we hypothesized the MMPI-3 SUI scale would demonstrate medium to large associations with suicidal experience and behaviors, future ideation, and interpersonal risk factors of suicide.

METHODS

A sample of 124 college students screened for elevated depressive symptoms completed a brief longitudinal study. Participants completed a baseline session including the MMPI-3 and criterion measures and three brief follow-ups every 2 weeks.

RESULTS

SUI scores were most robustly associated with increased risk for past suicidal ideation, planning, and perceived burdensomeness. Prospectively assessed suicidal ideation was also meaningfully associated with SUI. SUI scale elevations indicate an increased risk of suicide-related risk factors.

CONCLUSION

The MMPI-3 is a valuable tool to inform long-term suicide prevention for those experiencing elevated depressive symptoms as the SUI scale can assess past, current, and future suicide-related risk factors, including suicidal ideation and behaviors.

Sexual and Gender Minority Status and Suicide Mortality: An Explainable Artificial Intelligence Analysis

Objectives: Suicide risk is elevated in lesbian, gay, bisexual, and transgender (LGBT) individuals. Limited data on LGBT status in healthcare systems hinder our understanding of this risk. This study used natural language processing to extract LGBT status and a deep neural network (DNN) to examine suicidal death risk factors among US Veterans. Methods: Data on 8.8 million veterans with visits between 2010 and 2017 was used. A case-control study was performed, and suicide death risk was analyzed by a DNN. Feature impacts and interactions on the outcome were evaluated. Results: The crude suicide mortality rate was higher in LGBT patients. However, after adjusting for over 200 risk and protective factors, known LGBT status was associated with reduced risk compared to LGBT-Unknown status. Among LGBT patients, black, female, married, and older Veterans have a higher risk, while Veterans of various religions have a lower risk. Conclusion: Our results suggest that disclosed LGBT status is not directly associated with an increase suicide death risk, however, other factors (e.g., depression and anxiety caused by stigma) are associated with suicide death risks.

Psychometric properties and factor structure of the suicidal narrative inventory in major depression: A multicentric evaluation

BACKGROUND

The Suicidal Narrative Inventory (SNI) is a 38-item self-report measure developed to assess elements of the suicidal narrative, a subacute, predominantly cognitive, presuicidal construct. Our objectives were to assess the factor structure, validity, and reliability of the SNI-38 among adults with major depressive disorder (MDD).

METHODS

Using a cross-sectional design, we administered the Hindi version of the SNI along with other self-report measures to adults with MDD, recruited from 24 tertiary care hospitals across India. Confirmatory factor analysis (CFA) was performed to assess the factor structure of SNI-38. Reliability (internal consistency) was assessed using Cronbach's alpha (α). Convergent, discriminant, and criterion validity of the SNI-38 were tested by comparing it against other appropriate measures.

RESULTS

We collected usable responses from 654 Hindi-speaking participants (Mean age = 36.9 ± 11.9 years, 50.2% female). The eight-factor solution of the SNI showed good model fit indices (χ2[637] = 3345.58, p <.001, CFI =.98, and RMSEA =.08). Internal consistencies for the SNI subscale scores were good to excellent, α ranging from .73 to.92. While most subscales significantly converged with other measures, associations were comparatively weaker and inconsistent for the 'thwarted belongingness' and 'goal reengagement' subscales.

CONCLUSION

Consistent with prior data, our study confirmed an eight-factor solution and demonstrated adequate psychometric properties for the Hindi version of the SNI-38 in our sample. These findings provide empirical support for the use of SNI to assess the suicidal narrative among Indian adults with MDD.

Polygenic risk for suicide attempt is associated with lifetime suicide attempt in US soldiers independent of parental risk

BACKGROUND

Suicide is a leading cause of death worldwide. Whereas some studies have suggested that a direct measure of common genetic liability for suicide attempts (SA), captured by a polygenic risk score for SA (SA-PRS), explains risk independent of parental history, further confirmation would be useful. Even more unsettled is the extent to which SA-PRS is associated with lifetime non-suicidal self-injury (NSSI).

METHODS

We used summary statistics from the largest available GWAS study of SA to generate SA-PRS for two non-overlapping cohorts of soldiers of European ancestry. These were tested in multivariable models that included parental major depressive disorder (MDD) and parental SA.

RESULTS

In the first cohort, 417 (6.3 %) of 6573 soldiers reported lifetime SA and 1195 (18.2 %) reported lifetime NSSI. In a multivariable model that included parental history of MDD and parental history of SA, SA-PRS remained significantly associated with lifetime SA [aOR = 1.26, 95%CI:1.13-1.39, p < 0.001] per standardized unit SA-PRS]. In the second cohort, 204 (4.2 %) of 4900 soldiers reported lifetime SA, and 299 (6.1 %) reported lifetime NSSI. In a multivariable model that included parental history of MDD and parental history of SA, SA-PRS remained significantly associated with lifetime SA [aOR = 1.20, 95%CI:1.04-1.38, p = 0.014]. A combined analysis of both cohorts yielded similar results. In neither cohort or in the combined analysis was SA-PRS significantly associated with NSSI.

CONCLUSIONS

PRS for SA conveys information about likelihood of lifetime SA (but not NSSI, demonstrating specificity), independent of self-reported parental history of MDD and parental history of SA.

LIMITATIONS

At present, the magnitude of effects is small and would not be immediately useful for clinical decision-making or risk-stratified prevention initiatives, but this may be expected to improve with further iterations. Also critical will be the extension of these findings to more diverse populations.

Developing a risk prediction model for death at first suicide attempt-Identifying risk factors from Thailand's national suicide surveillance system data

More than 60% of suicides globally are estimated to take place in low- and middle-income nations. Prior research on suicide has indicated that over 50% of those who die by suicide do so on their first attempt. Nevertheless, there is a dearth of knowledge on the attributes of individuals who die on their first attempt and the factors that can predict mortality on the first attempt in these regions. The objective of this study was to create an individual-level risk-prediction model for mortality on the first suicide attempt. We analyzed records of individuals' first suicide attempts that occurred between May 1, 2017, and April 30, 2018, from the national suicide surveillance system, which includes all of the provinces of Thailand. Subsequently, a risk-prediction model for mortality on the first suicide attempt was constructed utilizing multivariable logistic regression and presented through a web-based application. The model's performance was assessed by calculating the area under the receiver operating curve (AUC), as well as measuring its sensitivity, specificity, and accuracy. Out of the 3,324 individuals who made their first suicide attempt, 50.5% of them died as a result of that effort. Nine out of the 21 potential predictors demonstrated the greatest predictive capability. These included male sex, age over 50 years old, unemployment, having a depressive disorder, having a psychotic illness, experiencing interpersonal problems such as being aggressively criticized or desiring plentiful attention, having suicidal intent, and displaying suicidal warning signals. The model demonstrated a good predictive capability, with an AUC of 0.902, a sensitivity of 84.65%, a specificity of 82.66%, and an accuracy of 83.63%. The implementation of this predictive model can assist physicians in conducting comprehensive evaluations of suicide risk in clinical settings and devising treatment plans for preventive intervention.

Explainable artificial intelligence models for predicting risk of suicide using health administrative data in Quebec

Suicide is a complex, multidimensional event, and a significant challenge for prevention globally. Artificial intelligence (AI) and machine learning (ML) have emerged to harness large-scale datasets to enhance risk detection. In order to trust and act upon the predictions made with ML, more intuitive user interfaces must be validated. Thus, Interpretable AI is one of the crucial directions which could allow policy and decision makers to make reasonable and data-driven decisions that can ultimately lead to better mental health services planning and suicide prevention. This research aimed to develop sex-specific ML models for predicting the population risk of suicide and to interpret the models. Data were from the Quebec Integrated Chronic Disease Surveillance System (QICDSS), covering up to 98% of the population in the province of Quebec and containing data for over 20,000 suicides between 2002 and 2019. We employed a case-control study design. Individuals were considered cases if they were aged 15+ and had died from suicide between January 1st, 2002, and December 31st, 2019 (n = 18339). Controls were a random sample of 1% of the Quebec population aged 15+ of each year, who were alive on December 31st of each year, from 2002 to 2019 (n = 1,307,370). We included 103 features, including individual, programmatic, systemic, and community factors, measured up to five years prior to the suicide events. We trained and then validated the sex-specific predictive risk model using supervised ML algorithms, including Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGBoost) and Multilayer perceptron (MLP). We computed operating characteristics, including sensitivity, specificity, and Positive Predictive Value (PPV). We then generated receiver operating characteristic (ROC) curves to predict suicides and calibration measures. For interpretability, Shapley Additive Explanations (SHAP) was used with the global explanation to determine how much the input features contribute to the models' output and the largest absolute coefficients. The best sensitivity was 0.38 with logistic regression for males and 0.47 with MLP for females; the XGBoost Classifier with 0.25 for males and 0.19 for females had the best precision (PPV). This study demonstrated the useful potential of explainable AI models as tools for decision-making and population-level suicide prevention actions. The ML models included individual, programmatic, systemic, and community levels variables available routinely to decision makers and planners in a public managed care system. Caution shall be exercised in the interpretation of variables associated in a predictive model since they are not causal, and other designs are required to establish the value of individual treatments. The next steps are to produce an intuitive user interface for decision makers, planners and other stakeholders like clinicians or representatives of families and people with live experience of suicidal behaviors or death by suicide. For example, how variations in the quality of local area primary care programs for depression or substance use disorders or increased in regional mental health and addiction budgets would lower suicide rates.

Predicting Lifetime Suicide Attempts in a Community Sample of Adolescents Using Machine Learning Algorithms

Suicide is a major global health concern and a prominent cause of death in adolescents. Previous research on suicide prediction has mainly focused on clinical or adult samples. To prevent suicides at an early stage, however, it is important to screen for risk factors in a community sample of adolescents. We compared the accuracy of logistic regressions, elastic net regressions, and gradient boosting machines in predicting suicide attempts by 17-year-olds in the Millennium Cohort Study (N = 7,347), combining a large set of self- and other-reported variables from different categories. Both machine learning algorithms outperformed logistic regressions and achieved similar balanced accuracies (.76 when using data 3 years before the self-reported lifetime suicide attempts and .85 when using data from the same measurement wave). We identified essential variables that should be considered when screening for suicidal behavior. Finally, we discuss the usefulness of complex machine learning models in suicide prediction.

Proximal characteristics of suicide attempts: a study in a public hospital in Spain

INTRODUCTION

Different parameters of suicide attempts treated since the implementation of the Attention to Suicide Risk Program (ARSUIC) in 2012 at the Hospital Ramón y Cajal in Madrid Region are described in this paper.

METHOD

The sample was composed of 107 patients and the information was collected through a questionnaire created ad hoc with the following variables: type of suicidal ideation; drug use immediately prior to the attempt; method (in case of drug overdosing: drug/s used); location; accessibility to rescue; planning; intentionality; criticism; and brakes.

RESULTS

Descriptive statistics were obtained and a comparison by gender was made through the χ2 and contingency coefficients tests. The data from the retrospective longitudinal study showed that the most common profile was of patients with unstructured ideas of death and no previous drug use who took an unplanned drug overdose in the family home, with the intention of self-harm or avoidance of discomfort, especially with benzodiazepines. Patients tend to ask for help afterwards and criticise the attempt, but potential restraints are often not recorded in the clinical report. Regarding the dissimilarities based on gender, statistically significant differences were found in prior alcohol consumption, in favour of men and in the overdose method, specifically with benzodiazepines, in favour of women.

CONCLUSIONS

Knowing the types of attempts at self-harm is essential for improving prevention, understanding and patient management.

Reduced attentional control in individuals with a history of suicide attempts compared to those with suicidal ideation: Results from a systematic review and meta-analysis

BACKGROUND

Neurocognitive profiles may be especially useful to identify factors that facilitate transitioning from contemplating suicide to attempting suicide. Generally, those who attempt suicide show greater disruptions in neurocognitive ability compared to those who think about suicide but do not proceed to attempt. The goal of this systematic review and meta-analysis is to test whether this pattern is observed with attentional control.

METHODS

We systematically searched PubMed, PsychINFO, CINAHL, and Google Scholar to find pertinent studies. All included studies compared attentional functioning using neutral stimuli. Each sample featured adults with a history of suicidal ideation (SI) and no history of suicide attempts (SA) compared to those with a history of SA.

RESULTS

We identified 15 studies with 32 effect sizes (N = 931; n = 506 with SI only; n = 425 with SA). SA groups, compared to SI groups, exhibited worse accuracy yet similar reaction time, suggesting a comparatively blunted speed-accuracy tradeoff. Relative to SI, SA groups performed worse on Stroop-like and Go/NoGo tasks. SA performed better than SI on Trail Making Test B, but not A.

LIMITATIONS

There were few available studies. Most samples were small. We did not differentiate current vs. past SI or high vs. low lethality SA. Only English and Spanish language articles were included.

CONCLUSIONS

Disrupted attentional control may convey risk for transitioning to SA from SI. More work is needed to determine which components of attention are most associated with suicide risk.

Predicting suicide death after emergency department visits with mental health or self-harm diagnoses

OBJECTIVE

Use health records data to predict suicide death following emergency department visits.

METHODS

Electronic health records and insurance claims from seven health systems were used to: identify emergency department visits with mental health or self-harm diagnoses by members aged 11 or older; extract approximately 2500 potential predictors including demographic, historical, and baseline clinical characteristics; and ascertain subsequent deaths by self-harm. Logistic regression with lasso and random forest models predicted self-harm death over 90 days after each visit.

RESULTS

Records identified 2,069,170 eligible visits, 899 followed by suicide death within 90 days. The best-fitting logistic regression with lasso model yielded an area under the receiver operating curve of 0.823 (95% CI 0.810-0.836). Visits above the 95th percentile of predicted risk included 34.8% (95% CI 31.1-38.7) of subsequent suicide deaths and had a 0.303% (95% CI 0.261-0.346) suicide death rate over the following 90 days. Model performance was similar across subgroups defined by age, sex, race, and ethnicity.

CONCLUSIONS

Machine learning models using coded data from health records have moderate performance in predicting suicide death following emergency department visits for mental health or self-harm diagnosis and could be used to identify patients needing more systematic follow-up.

Comorbidity profiles in older patients last seen by mental health prior to suicide attempt

OBJECTIVES

Suicide in late life is a public health concern. Determining profiles of psychiatric/medical comorbidity in those who attempt while engaged in mental health services may assist with prevention. We identified comorbidity profiles and their association with utilization, means, and fatality in a national sample who attempted suicide.

METHODS

Using latent class analysis, all patients aged ≥ 65 from the Department of Veterans Affairs (VA) healthcare services (2012-2018) last seen in mental health prior to suicide attempt were included. Diagnoses and attempt data were obtained from VA and Center for Medicare & Medicaid Services, VA Suicide Prevention Applications Network, and VA National Mortality Data Repository.

RESULTS

2,269 patients were clustered into three profiles, all with high probability of depression. Profiles included minimal comorbidity (50.4%), high medical comorbidity (28.6%), and high (psychiatric/medical) comorbidity (21.0%). Over half (61.7%) attempted suicide within one week of their visit. The class with highest comorbidity had lowest proportion of fatal attempts, while minimal comorbidity class had highest proportion.

CONCLUSIONS

Older patients last seen in mental health prior to suicide attempt were characterized by depression and varying additional comorbidity and attempt-related factors. Findings have implications for risk assessment and intervention in mental health settings, beyond depression.

Mental health of Brazilian physicians: a nationwide crosssectional study to investigate factors associated with the prevalence of suicide plans and attempts

OBJECTIVE

To report on suicide plans and attempts among Brazilian physicians and to investigate the associated risk factors.

METHODS

From January 2018 to January 2019, a nationwide online survey was conducted among Brazilian physicians using the Tool for Assessment of Suicide Risk and the Satisfaction with Life Scale. Multivariate exploratory associations of demographic, psychological, and work-related factors were performed on reports of suicide plans and attempts.

RESULTS

Of the 4,148 participants, 1,946 (53.5%) were male, 2,527 (60.9%) were 30 to 60 years old, 2,675 (64.5%) had two to four jobs, and 1,725 (41.6%) worked 40 to 60 hours a week. The overall prevalence of suicide plans was 8.8% (n=364), and suicide attempts were reported by 3.2% (n=133) of participants. Daily emotional exhaustion (ORadj = 7.857; 95%CI 2.282-27.051, p = 0.002), weekly emotional exhaustion (ORadj = 7.953; 95%CI 2.403-26.324, p = 0.001), daily frustration at work (ORadj = 3.093; 95%CI 1.711-5.588, p < 0.001), and bisexuality (ORadj = 5.083; 95%CI 2.544-10.158, p < 0.001) were significantly associated with higher odds of suicide. Extremely dissatisfied physicians reported suicide plans and attempts in 38.3% of cases, whereas extremely satisfied physicians reported suicide plans and attempts in only 2.8% of cases (p < 0.001).

CONCLUSION

Brazilian physicians with a history of suicide plans and attempts express emotional exhaustion and frustration at work. There is an urgent need for actions to promote professional safeguards and resilience.

A Randomized Control Trial of a Digital Health Tool for Safer Firearm and Medication Storage for Patients with Suicide Risk

Most patients with suicide risk do not receive recommendations to reduce access to lethal means due to a variety of barriers (e.g., lack of provider time, training). Determine if highly efficient population-based EHR messaging to visit the Lock to Live (L2L) decision aid impacts patient-reported storage behaviors. Randomized trial. Integrated health care system serving Denver, CO. Served by primary care or mental health specialty clinic in the 75-99.5th risk percentile on a suicide attempt or death prediction model. Lock to Live (L2L) is a web-based decision aid that incorporates patients' values into recommendations for safe storage of lethal means, including firearms and medications. Anonymous survey that determined readiness to change: pre-contemplative (do not believe in safe storage), contemplative (believe in safe storage but not doing it), preparation (planning storage changes) or action (safely storing). There were 21,131 patients randomized over a 6-month period with a 27% survey response rate. Many (44%) had access to a firearm, but most of these (81%) did not use any safe firearm storage behaviors. Intervention patients were more likely to be categorized as preparation or action compared to controls for firearm storage (OR = 1.30 (1.07-1.58)). When examining action alone, there were no group differences. There were no statistically significant differences for any medication storage behaviors. Selection bias in those who responded to survey. Efficiently sending an EHR invitation message to visit L2L encouraged patients with suicide risk to consider safer firearm storage practices, but a stronger intervention is needed to change storage behaviors. Future studies should evaluate whether combining EHR messaging with provider nudges (e.g., brief clinician counseling) changes storage behavior.ClinicalTrials.gov: NCT05288517.

A Moderation Model for Bolstering Resilience to Suicidal Psychopathology: Positive Sociopsychological Constructs and Coping Flexibilities Buffering the Impact of Daily Life Stress Among Medical Students

ABSTRACT

Positive sociopsychological constructs and coping flexibility may be helpful for alleviating suicidal psychopathology, although relatively little research has examined this possibility among medical students. This survey aimed to investigate whether positive sociopsychological constructs and coping flexibility could buffer the negative impact of daily life stress and bolster the resilience to suicidal ideations and attempts among medical students. This cross-sectional model was based on a study of 787 Indian medical students (725 women and 62 men; mean age, 21.08 years; SD = 2.78; range, 19-37 years) who were asked to complete a battery of self-administered questionnaires. For the purpose of determining the independent and interaction impacts of potential variables of influence, hierarchical multiple linear regression models were used. The moderation analysis investigated that the association between daily life stress and suicidal ideation was buffered among the students having higher levels of positive mental health and coping flexibility, whereas this association was no longer significant at the highest level of positive mental health. Furthermore, the relationship between daily life stress and suicidal attempts continued to be buffered by above-average levels of coping flexibility, emotional stability (ES), and optimism. These findings represent that promoting positive mental health, coping flexibility, ES, and optimism may be a promising approach to mitigate suicidal thoughts and attempts in interventions for medical students at high risk. These modifiable moderating factors can be enhanced by empirically supported treatment and prevention efforts to bolster suicidal resilience.

Maladaptive daydreaming: A shortened assessment measure and its mental health correlates in a large United States sample

BACKGROUND AND AIMS

Maladaptive Daydreaming (MD) is a suggested syndrome where individuals addictively engage in fanciful, narrative and emotional daydreaming for hours on end, often relying on stereotypical movements and music to facilitate the absorbed state. Many individuals suffering from MD to the point of clinically significant distress and functional impairment have advocated for its medicalization as a disorder. Maladaptive daydreamers exhibit high rates of psychopathology, but most studies were biased by self-selection. We developed a brief measure for efficient assessment of suspected MD and then administered it in a large non-selected US sample to gauge the significance of MD for public mental health.

METHODS

Two previous datasets were utilized to develop the 5-item measure, labeled the Maladaptive Daydreaming Short Form (MD-SF5). Then, a large survey was conducted using the Qualtrics panel, administering the MD-SF5 alongside several validated measures of mental health to a general sample of panelists (N = 2512, 84.6% females, age M = 39.74, SD = 18.53, Race/Ethnicity: 66.3% White, 14.7% Black, 9.3% Hispanic, and 9.7% Other).

RESULTS

The MD-SF5 showed good to excellent agreement with the existing measure. Generally, the new sample had high psychopathology rates. Suspected MD was associated with psychological distress, loneliness, psychotic experiences, heavy drinking, and suicidality. Notably, even after controlling for psychological distress, suspected maladaptive daydreamers were more than twice as likely to have recently attempted suicide (Odds Ratio = 2.44, 95% CI [1.44, 4.16], Wald = 10.86, p = .001).

DISCUSSION AND CONCLUSIONS

MD harbors public health significance and can be screened for with a short self-report tool. Thus, MD should be addressed by mental health practitioners.

Scoping Review of Postvention for Mental Health Providers Following Patient Suicide

INTRODUCTION

As suicides among military personnel continue to climb, we sought to determine best practices for supporting military mental health clinicians following patient suicide loss (i.e., postvention).

MATERIALS AND METHODS

We conducted a scoping review of the literature using Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews guidelines. Our initial search of academic databases generated 2,374 studies, of which 122 were included in our final review. We categorized postvention recommendations based on the socioecological model (i.e., recommendations at the individual provider, supervisory/managerial, organizational, and discipline levels) and analyzed them using a narrative synthesizing approach.

RESULTS

Extracted recommendations (N = 358) comprised those at the provider (n = 94), supervisory/managerial (n = 90), organization (n = 105), and discipline (n = 69) levels.

CONCLUSIONS

The literature converges on the need for formal postvention protocols that prioritize (1) training and education and (2) emotional and instrumental support for the clinician. Based on the scoped literature, we propose a simple postvention model for military mental health clinicians and recommend a controlled trial testing of its effectiveness.

Factor structure, reliability, and validity of the revised Suicide Crisis Inventory in major depression: A multicentric Indian study

BACKGROUND

The revised Suicide Crisis Inventory (SCI)-2 is a self-report measure to assess the suicide crisis syndrome (SCS). We aimed to assess the factor structure, reliability, and validity of SCI-2 among adults with major depression.

METHODS

Using a cross-sectional design, between November 2021 and August 2022, the Hindi SCI-2, along with other self-report measures, was administered to Indian adult respondents clinically diagnosed with major depression across 24 centers in India. Confirmatory factor analysis was carried out to test the factor structure of SCI-2. Additionally, convergent, discriminant, and criterion validity were tested using bivariate or biserial correlations, as appropriate.

RESULTS

We obtained responses from 654 participants (Mean age = 36.9 ± 11.9 years, 50.2 % female). The SCI-2 fit both a one-factor (χ2[1769] = 14,150.74, p < .001, CFI = 0.98, RMSEA = 0.10), and five-factor solution (χ2[1759] = 13,130.83, p < .001,CFI = 0.98, RMSEA = 0.10) with the five-factor solution providing a significantly better fit. Internal consistencies of the SCI-2 total and subscale scores ranged from good to excellent. Most subscales significantly converged with each other and with other relevant measures although these associations were weak for thwarted belongingness and goal reengagement subscales. Small to moderate associations were noted in support of discriminant and criterion validity.

LIMITATIONS

We could not assess the predictive validity of SCI-2 for suicidal behaviors.

CONCLUSION

Consistent with prior data, the Hindi SCI-2 fit a five-factor solution and showed good psychometric properties. These findings support the use of SCI-2 to assess SCS among Indian adults with major depression.

Determining Distinct Suicide Attempts From Recurrent Electronic Health Record Codes: Classification Study

BACKGROUND

Prior suicide attempts are a relatively strong risk factor for future suicide attempts. There is growing interest in using longitudinal electronic health record (EHR) data to derive statistical risk prediction models for future suicide attempts and other suicidal behavior outcomes. However, model performance may be inflated by a largely unrecognized form of "data leakage" during model training: diagnostic codes for suicide attempt outcomes may refer to prior attempts that are also included in the model as predictors.

OBJECTIVE

We aimed to develop an automated rule for determining when documented suicide attempt diagnostic codes identify distinct suicide attempt events.

METHODS

From a large health care system's EHR, we randomly sampled suicide attempt codes for 300 patients with at least one pair of suicide attempt codes documented at least one but no more than 90 days apart. Supervised chart reviewers assigned the clinical settings (ie, emergency department [ED] versus non-ED), methods of suicide attempt, and intercode interval (number of days). The probability (or positive predictive value) that the second suicide attempt code in a given pair of codes referred to a distinct suicide attempt event from its preceding suicide attempt code was calculated by clinical setting, method, and intercode interval.

RESULTS

Of 1015 code pairs reviewed, 835 (82.3%) were nonindependent (ie, the 2 codes referred to the same suicide attempt event). When the second code in a pair was documented in a clinical setting other than the ED, it represented a distinct suicide attempt 3.3% of the time. The more time elapsed between codes, the more likely the second code in a pair referred to a distinct suicide attempt event from its preceding code. Code pairs in which the second suicide attempt code was assigned in an ED at least 5 days after its preceding suicide attempt code had a positive predictive value of 0.90.

CONCLUSIONS

EHR-based suicide risk prediction models that include International Classification of Diseases codes for prior suicide attempts as a predictor may be highly susceptible to bias due to data leakage in model training. We derived a simple rule to distinguish codes that reflect new, independent suicide attempts: suicide attempt codes documented in an ED setting at least 5 days after a preceding suicide attempt code can be confidently treated as new events in EHR-based suicide risk prediction models. This rule has the potential to minimize upward bias in model performance when prior suicide attempts are included as predictors in EHR-based suicide risk prediction models.

Predicting risk of suicidal behavior from insurance claims data vs. linked data from insurance claims and electronic health records

PURPOSE

Observational studies assessing effects of medical products on suicidal behavior often rely on health record data to account for pre-existing risk. We assess whether high-dimensional models predicting suicide risk using data derived from insurance claims and electronic health records (EHRs) are superior to models using data from insurance claims alone.

METHODS

Data were from seven large health systems identified outpatient mental health visits by patients aged 11 or older between 1/1/2009 and 9/30/2017. Data for the 5 years prior to each visit identified potential predictors of suicidal behavior typically available from insurance claims (e.g., mental health diagnoses, procedure codes, medication dispensings) and additional potential predictors available from EHRs (self-reported race and ethnicity, responses to Patient Health Questionnaire or PHQ-9 depression questionnaires). Nonfatal self-harm events following each visit were identified from insurance claims data and fatal self-harm events were identified by linkage to state mortality records. Random forest models predicting nonfatal or fatal self-harm over 90 days following each visit were developed in a 70% random sample of visits and validated in a held-out sample of 30%. Performance of models using linked claims and EHR data was compared to models using claims data only.

RESULTS

Among 15 845 047 encounters by 1 574 612 patients, 99 098 (0.6%) were followed by a self-harm event within 90 days. Overall classification performance did not differ between the best-fitting model using all data (area under the receiver operating curve or AUC = 0.846, 95% CI 0.839-0.854) and the best-fitting model limited to data available from insurance claims (AUC = 0.846, 95% CI 0.838-0.853). Competing models showed similar classification performance across a range of cut-points and similar calibration performance across a range of risk strata. Results were similar when the sample was limited to health systems and time periods where PHQ-9 depression questionnaires were recorded more frequently.

CONCLUSION

Investigators using health record data to account for pre-existing risk in observational studies of suicidal behavior need not limit that research to databases including linked EHR data.

Assessment and management of individuals consulting for a suicidal crisis: A European Delphi method-based consensus guidelines

INTRODUCTION

Managing patient with suicidal thoughts and behaviours presents significant challenges due to the scarcity of robust evidence and clear guidance. This study sought to develop a comprehensive set of practical guidelines for the assessment and management of suicidal crises.

MATERIALS AND METHODS

Utilizing the Delphi methodology, 80 suicide clinician and research experts agreed on a series of recommendations. The process involved two iterative rounds of surveys to assess agreement with drafted recommendations, inviting panellists to comment and vote, culminating in 43 consensus recommendations approved with at least 67% agreement. These consensus recommendations fall into three main categories: clinical assessment, immediate care, and long-term approaches.

RESULTS

The panel formulated 43 recommendations spanning suicidal crisis recognition to continuous long-term care. These guidelines underscore systematic proactive suicide risk screening, in-depth medical and toxicological assessment, and suicide risk appraisal considering personal, clinical factors and collateral information from family. The immediate care directives emphasize a secure environment, continuous risk surveillance, collaborative decision-making, including potential hospitalization, sensible pharmacological management, safety planning, and lethal means restriction counselling. Every discharge should be accompanied by prompt follow-up care incorporating proactive case management and multi-modal approach involving crisis lines, brief contact, and psychotherapeutic and pharmacological interventions.

CONCLUSIONS

This study generated comprehensive guidelines addressing care for individuals in suicidal crises, covering pre- to post-discharge care. These practical recommendations can guide clinicians in managing patients with suicidal thoughts and behaviours, improve patient safety, and ultimately contribute to the prevention of future suicidal crises.

Factors associated with suicide attempts in the antecedent illness trajectory of bipolar disorder and schizophrenia

BACKGROUND

Factors associated with suicide attempts during the antecedent illness trajectory of bipolar disorder (BD) and schizophrenia (SZ) are poorly understood.

METHODS

Utilizing the Rochester Epidemiology Project, individuals born after 1985 in Olmsted County, MN, presented with first episode mania (FEM) or psychosis (FEP), subsequently diagnosed with BD or SZ were identified. Patient demographics, suicidal ideation with plan, self-harm, suicide attempts, psychiatric hospitalizations, substance use, and childhood adversities were quantified using the electronic health record. Analyses pooled BD and SZ groups with a transdiagnostic approach given the two diseases were not yet differentiated. Factors associated with suicide attempts were examined using bivariate methods and multivariable logistic regression modeling.

RESULTS

A total of 205 individuals with FEM or FEP (BD = 74, SZ = 131) were included. Suicide attempts were identified in 39 (19%) patients. Those with suicide attempts during antecedent illness trajectory were more likely to be female, victims of domestic violence or bullying behavior, and have higher rates of psychiatric hospitalizations, suicidal ideation with plan and/or self-harm, as well as alcohol, drug, and nicotine use before FEM/FEP onset. Based on multivariable logistic regression, three factors remained independently associated with suicidal attempts: psychiatric hospitalization (OR = 5.84, 95% CI 2.09-16.33, p < 0.001), self-harm (OR = 3.46, 95% CI 1.29-9.30, p = 0.014), and nicotine use (OR = 3.02, 95% CI 1.17-7.76, p = 0.022).

CONCLUSION

Suicidal attempts were prevalent during the antecedents of BD and SZ and were associated with several risk factors before FEM/FEP. Their clinical recognition could contribute to improve early prediction and prevention of suicide during the antecedent illness trajectory of BD and SZ.

EEG wakefulness regulation in transdiagnostic patients after a recent suicide attempt

OBJECTIVE

Decades of research have not yet produced statistically reliable predictors of preparatory behavior eventually leading to suicide attempts or deaths by suicide. As the nature of suicidal behavior is complex, it is best investigated in a transdiagnostic approach, while assessing objective markers, as proposed by the Research Domain Criteria (Cuthbert, 2013).

METHODS

A 15-min resting-state EEG was recorded in 45 healthy controls, and 49 transdiagnostic in-patients with a recent (<6 months) suicide attempt. Brain arousal regulation in eyes-closed condition was assessed with the Vigilance Algorithm Leipzig (VIGALL) (Sander et al., 2015).

RESULTS

A significant incline of median vigilance and vigilance slope was observed in patients within the first 3-min of the EEG recording. Additionally, a significant positive correlation of self-reported suicidal ideation with the vigilance slope over 15-min recording time, as well as a significant negative correlation with EEG vigilance stage A1 during the first 3-min was found.

CONCLUSIONS

Transdiagnostic patients with a recent suicide attempt show a distinct vigilance regulation pattern. Further studies including a control group consisting of patients without life-time suicide attempts are needed to increase the clinical utility of the findings.

SIGNIFICANCE

These findings might serve as potential objective markers of suicidal behavior.

The Ethics of Risk Prediction for Psychosis and Suicide Attempt in Youth Mental Health

OBJECTIVE

To identify potential clinical utility of polygenic risk scores (PRS) and exposomic risk scores (ERS) for psychosis and suicide attempt in youth and assess the ethical implications of these tools.

STUDY DESIGN

We conducted a narrative literature review of emerging findings on PRS and ERS for suicide and psychosis as well as a literature review on the ethics of PRS. We discuss the ethical implications of the emerging findings for the clinical potential of PRS and ERS.

RESULTS

Emerging evidence suggests that PRS and ERS may offer clinical utility in the relatively near future but that this utility will be limited to specific, narrow clinical questions, in contrast to the suggestion that population-level screening will have sweeping impact. Combining PRS and ERS might optimize prediction. This clinical utility would change the risk-benefit balance of PRS, and further empirical assessment of proposed risks would be necessary. Some concerns for PRS, such as those about counseling, privacy, and inequities, apply to ERS. ERS raise distinct ethical challenges as well, including some that involve informed consent and direct-to-consumer advertising. Both raise questions about the ethics of machine-learning/artificial intelligence approaches.

CONCLUSIONS

Predictive analytics using PRS and ERS may soon play a role in youth mental health settings. Our findings help educate clinicians about potential capabilities, limitations, and ethical implications of these tools. We suggest that a broader discussion with the public is needed to avoid overenthusiasm and determine regulations and guidelines for use of predictive scores.

External Validation and Updating of a Statistical Civilian-Based Suicide Risk Model in US Naval Primary Care

Importance

Suicide remains an ongoing concern in the US military. Statistical models have not been broadly disseminated for US Navy service members.

Objective

To externally validate and update a statistical suicide risk model initially developed in a civilian setting with an emphasis on primary care.

Design, Setting, and Participants

This retrospective cohort study used data collected from 2007 through 2017 among active-duty US Navy service members. The external civilian model was applied to every visit at Naval Medical Center Portsmouth (NMCP), its NMCP Naval Branch Health Clinics (NBHCs), and TRICARE Prime Clinics (TPCs) that fall within the NMCP area. The model was retrained and recalibrated using visits to NBHCs and TPCs and updated using Department of Defense (DoD)-specific billing codes and demographic characteristics, including expanded race and ethnicity categories. Domain and temporal analyses were performed with bootstrap validation. Data analysis was performed from September 2020 to December 2022.

Exposure

Visit to US NMCP.

Main Outcomes and Measures

Recorded suicidal behavior on the day of or within 30 days of a visit. Performance was assessed using area under the receiver operating curve (AUROC), area under the precision recall curve (AUPRC), Brier score, and Spiegelhalter z-test statistic.

Results

Of the 260 583 service members, 6529 (2.5%) had a recorded suicidal behavior, 206 412 (79.2%) were male; 104 835 (40.2%) were aged 20 to 24 years; and 9458 (3.6%) were Asian, 56 715 (21.8%) were Black or African American, and 158 277 (60.7%) were White. Applying the civilian-trained model resulted in an AUROC of 0.77 (95% CI, 0.74-0.79) and an AUPRC of 0.004 (95% CI, 0.003-0.005) at NBHCs with poor calibration (Spiegelhalter P < .001). Retraining the algorithm improved AUROC to 0.92 (95% CI, 0.91-0.93) and AUPRC to 0.66 (95% CI, 0.63-0.68). Number needed to screen in the top risk tiers was 366 for the external model and 200 for the retrained model; the lower number indicates better performance. Domain validation showed AUROC of 0.90 (95% CI, 0.90-0.91) and AUPRC of 0.01 (95% CI, 0.01-0.01), and temporal validation showed AUROC of 0.75 (95% CI, 0.72-0.78) and AUPRC of 0.003 (95% CI, 0.003-0.005).

Conclusions and Relevance

In this cohort study of active-duty Navy service members, a civilian suicide attempt risk model was externally validated. Retraining and updating with DoD-specific variables improved performance. Domain and temporal validation results were similar to external validation, suggesting that implementing an external model in US Navy primary care clinics may bypass the need for costly internal development and expedite the automation of suicide prevention in these clinics.

Quantitative content analysis of Freedom of Information requests examining the extent and variations of tools and training for conducting suicide risk assessments in NHS Trusts across England

OBJECTIVES

Determining the risk for suicide is a difficult endeavour. Clinical guidance in the UK explicitly advises against using risk assessment tools and scales to determine suicide risk. Based on Freedom of Information (FoI) requests made to NHS Trusts in England, this study provides an overview of suicide risk assessment tools in use, training provided in how to use such assessments, and explores implementation of suicide risk assessment guidance in practice in English NHS Trusts.

DESIGN

A cross-sectional survey of suicide risk assessment tools and training gathered via FoI requests and subjected to a content analysis.

SETTING

FoI requests were submitted to NHS Trusts across England.

RESULTS

A wide variety of suicide risk assessments tools were identified as being used in practice, with several trusts reported using more than one tool to determine suicide risk. Forty-one trusts reported using locally developed, unvalidated, tools to assess risk of suicide and 18 stated they do not use a tool. Ten trusts stated they do not train their staff in suicide risk assessment while 13 reported use of specific suicide risk assessment training. Sixty-two trusts stated they do not centrally record the number of assessments conducted or how many individuals are identified as at risk. Content analysis indicated the frequent wider assessment of risk not restricted to suicide risk.

CONCLUSIONS

There is wide variation in suicide risk assessment tools being used in practice and some lack of specific training for healthcare staff in determining suicide risk. Few trusts routinely record the number of assessments being conducted or the number of individuals identified at high risk. Implementation of specific training is necessary for the suicide risk assessment process to identify patient needs and develop therapeutic engagement. Routinely recording how many assessments are conducted is a crucial step in improving suicide prevention.

Analytical reference framework to analyze non-COVID-19 events

BACKGROUND

The COVID-19 pandemic has disrupted the healthcare system, leading to delays in detection of other non-COVID-19 diseases. This paper presents ANE Framework (Analytics for Non-COVID-19 Events), a reliable and user-friendly analytical forecasting framework designed to predict the number of patients with non-COVID-19 diseases. Prior to 2020, there were analytical models focused on specific illnesses and contexts. Then, most models have focused on understanding COVID-19 behavior. There is a lack of analytical frameworks that enable disease forecasting for non-COVID-19 diseases.

METHODS

The ANE Framework utilizes time series analysis to generate forecasting models. The framework leverages daily data from official government sources and employs SARIMA models to forecast the number of non-COVID-19 cases, such as tuberculosis and suicide attempts.

RESULTS

The framework was tested on five different non-COVID-19 events. The framework performs well across all events, including tuberculosis and suicide attempts, with a Mean Absolute Percentage Error (MAPE) of up to 20% and the consistency remains independent of the behavior of each event. Moreover, a pairwise comparison of averages can lead to over or underestimation of the impact. The disruption caused by the pandemic resulted in a 17% gap (2383 cases) between expected and reported tuberculosis cases, and a 19% gap (2464 cases) for suicide attempts. These gaps varied between 20 and 64% across different cities and regions. The ANE Framework has proven to be reliable for analyzing several diseases and exhibits the flexibility to incorporate new data from various sources. Regular updates and the inclusion of new associated data enhance the framework's effectiveness.

CONCLUSIONS

Current pandemic shows the necessity of developing flexible models to be adapted to different illness data. The framework developed proved to be reliable for the different diseases analyzed, presenting enough flexibility to update with new data or even include new data from different databases. To keep updated on the result of the project allows the inclusion of new data associated with it. Similarly, the proposed strategy in the ANE framework allows for improving the quality of the obtained results with news events.

Toward objective characterizations of suicide risk: A narrative review of laboratory-based cognitive and behavioral tasks

Although suicide is a leading cause of preventable death worldwide, current prevention efforts have failed to substantively mitigate suicide risk. Suicide research has traditionally relied on subjective reports that may not accurately differentiate those at high versus minimal risk. This narrative review supports the inclusion of objective task-based measures in suicide research to complement existing subjective batteries. The article: 1) outlines risk factors proposed by contemporary theories of suicide and highlights recent empirical findings supporting these theories; 2) discusses ongoing challenges associated with current risk assessment tools and their ability to accurately evaluate risk factors; and 3) analyzes objective laboratory measures that can be implemented alongside traditional measures to enhance the precision of risk assessment. To illustrate the potential of these methods to improve our understanding of suicide risk, the article reviews how acute stress responses in a laboratory setting can be modeled, given that stress is a major precipitant for suicidal behavior. More precise risk assessment strategies can emerge if objective measures are implemented in conjunction with traditional subjective measures.

Predicting risk of suicidal behaviour after initiation of selective serotonin reuptake inhibitors in children, adolescents and young adults: protocol for development and validation of clinical prediction models

INTRODUCTION

There is concern regarding suicidal behaviour risk during selective serotonin reuptake inhibitor (SSRI) treatment among the young. A clinically useful model for predicting suicidal behaviour risk should have high predictive performance in terms of discrimination and calibration; transparency and ease of implementation are desirable.

METHODS AND ANALYSIS

Using Swedish national registers, we will identify individuals initiating an SSRI aged 8-24 years 2007-2020. We will develop: (A) a model based on a broad set of predictors, and (B) a model based on a restricted set of predictors. For the broad predictor model, we will consider an ensemble of four base models: XGBoost (XG), neural net (NN), elastic net logistic regression (EN) and support vector machine (SVM). The predictors with the greatest contribution to predictive performance in the base models will be determined. For the restricted predictor model, clinical input will be used to select predictors based on the top predictors in the broad model, and inputted in each of the XG, NN, EN and SVM models. If any show superiority in predictive performance as defined by the area under the receiver-operator curve, this model will be selected as the final model; otherwise, the EN model will be selected. The training and testing samples will consist of data from 2007 to 2017 and from 2018 to 2020, respectively. We will additionally assess the final model performance in individuals receiving a depression diagnosis within 90 days before SSRI initiation.The aims are to (A) develop a model predicting suicidal behaviour risk after SSRI initiation among children and youths, using machine learning methods, and (B) develop a model with a restricted set of predictors, favouring transparency and scalability.

ETHICS AND DISSEMINATION

The research is approved by the Swedish Ethical Review Authority (2020-06540). We will disseminate findings by publishing in peer-reviewed open-access journals, and presenting at international conferences.

Leveraging Natural Language Processing to Improve Electronic Health Record Suicide Risk Prediction for Veterans Health Administration Users

Background: Suicide risk prediction models frequently rely on structured electronic health record (EHR) data, including patient demographics and health care usage variables. Unstructured EHR data, such as clinical notes, may improve predictive accuracy by allowing access to detailed information that does not exist in structured data fields. To assess comparative benefits of including unstructured data, we developed a large case-control dataset matched on a state-of-the-art structured EHR suicide risk algorithm, utilized natural language processing (NLP) to derive a clinical note predictive model, and evaluated to what extent this model provided predictive accuracy over and above existing predictive thresholds.

Methods: We developed a matched case-control sample of Veterans Health Administration (VHA) patients in 2017 and 2018. Each case (all patients that died by suicide in that interval, n = 4,584) was matched with 5 controls (patients who remained alive during treatment year) who shared the same suicide risk percentile. All sample EHR notes were selected and abstracted using NLP methods. We applied machine-learning classification algorithms to NLP output to develop predictive models. We calculated area under the curve (AUC) and suicide risk concentration to evaluate predictive accuracy overall and for high-risk patients.

Results: The best performing NLP-derived models provided 19% overall additional predictive accuracy (AUC = 0.69; 95% CI, 0.67, 0.72) and 6-fold additional risk concentration for patients at the highest risk tier (top 0.1%), relative to the structured EHR model.

Conclusions: The NLP-supplemented predictive models provided considerable benefit when compared to conventional structured EHR models. Results support future structured and unstructured EHR risk model integrations.

Predictive value of suicidal risk assessment using data from China's largest suicide prevention hotline

BACKGROUND

Suicide hotlines are widely used, with potential for identification of callers at especially high risk.

METHODS

This prospective study was conducted at the largest psychological support hotline in China. From 2015 to 2017, all distressed callers were consecutively included and assessed, using a standardized scale consisting of 12 elements, yielding scores of high risk (8-16), moderate risk (4-7), and low risk (0-3) for suicidal act. All high-risk and half of moderate- and low-risk callers were scheduled for a 12-month follow-up. Main outcomes were suicidal acts (nonlethal attempt, death) over follow-up.

RESULTS

Of 21,346 fully assessed callers, 5822, 11,791, and 3733 were classified as high-, moderate-, or low-risk for suicidal acts, with 8869 callers (4076 high-, 3258 moderate-, and 1535 low-risk) followed up over 12 months. Over follow-up, 802 (9.0 %) callers attempted suicide or died by suicide. The high-risk callers (15.1 %) had 3-fold higher risk for subsequent suicidal acts than moderate- (5.1 %) and 12-fold higher risk than low-risk callers (1.3 %). The weighted sensitivity, specificity, and positive predictive value of high risk scores were 56.4 %, 74.9 %, and 14.4 %.

LIMITATIONS

Assessed callers with different risk levels were followed disproportionally.

CONCLUSIONS

Suicidal risk assessment during a hotline call is both feasible and predictive of risk, guiding resource allocation to higher risk callers.

Risk of death by suicide following self-harm presentations to healthcare: development and validation of a multivariable clinical prediction rule (OxSATS)

BACKGROUND

Assessment of suicide risk in individuals who have self-harmed is common in emergency departments, but is often based on tools developed for other purposes.

OBJECTIVE

We developed and validated a predictive model for suicide following self-harm.

METHODS

We used data from Swedish population-based registers. A cohort of 53 172 individuals aged 10+ years, with healthcare episodes of self-harm, was split into development (37 523 individuals, of whom 391 died from suicide within 12 months) and validation (15 649 individuals, 178 suicides within 12 months) samples. We fitted a multivariable accelerated failure time model for the association between risk factors and time to suicide. The final model contains 11 factors: age, sex, and variables related to substance misuse, mental health and treatment, and history of self-harm. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis guidelines were followed for the design and reporting of this work.

FINDINGS

An 11-item risk model to predict suicide was developed using sociodemographic and clinical risk factors, and showed good discrimination (c-index 0.77, 95% CI 0.75 to 0.78) and calibration in external validation. For risk of suicide within 12 months, using a 1% cut-off, sensitivity was 82% (75% to 87%) and specificity was 54% (53% to 55%). A web-based risk calculator is available (Oxford Suicide Assessment Tool for Self-harm or OxSATS).

CONCLUSIONS

OxSATS accurately predicts 12-month risk of suicide. Further validations and linkage to effective interventions are required to examine clinical utility.

CLINICAL IMPLICATIONS

Using a clinical prediction score may assist clinical decision-making and resource allocation.

Associations between nonfatal firearm injuries and risk of subsequent suicide among Veteran VA users: A retrospective cohort study

BACKGROUND

Suicide is a leading cause of death in the United States, particularly among Veterans. Nonfatal firearm injuries may indicate subsequent risk of suicide and, thus, provide important opportunities for prevention in emergency departments and other health care settings. We used a retrospective cohort design to analyze associations between nonfatal firearm injuries and subsequent suicide among all Veterans who used U.S. Department of Veterans Affairs (VA) health care, nationally, between 2010 and 2019.

METHODS

We linked VA health care and mortality data to identify VA users, nonfatal firearm injuries, and deaths. International Classification of Diseases (ICD)-10th Revision cause-of-death codes were used to identify suicides. Veterans' firearm injuries and their intent were categorized using cause-of-injury codes from the ICD Clinical Modification-9th and 10th Revisions systems. Using bivariable and multivariable regression, we estimated risk of subsequent suicide among Veterans with, versus without, nonfatal firearm injuries. Among Veterans with nonfatal firearm injuries, we examined characteristics associated with subsequent suicide; electronic health record (chart) reviews explored documentation about firearm access among those who died.

RESULTS

Among 9,817,020 VA-using Veterans, 11,503 experienced nonfatal firearm injuries (64.9% unintentional, 12.3% intentional self-harm, 18.5% assault). Of these, 69 (0.6%) subsequently died by suicide (42 involving firearms). The odds of subsequent suicide among Veterans with, versus without, nonfatal firearm injuries were 2.4 (95% confidence interval 1.9-3.0); odds were only slightly attenuated in multivariable modeling. Among Veterans with nonfatal firearm injuries, those with depression or substance use disorder diagnoses had twice the odds of subsequent suicide than those without. Chart reviews identified small proportions of suicide decedents who were assessed for (21.7%), and/or counseled about (15.9%), firearm access.

CONCLUSIONS

Findings suggest that Veterans' nonfatal firearm injuries, regardless of injury intent, may be important but underutilized opportunities for suicide prevention. Future work should explore mechanisms to reduce risk among these patients.

Identifying and mitigating moral injury risks in military behavioral health providers

The term "moral injury" was initially used to describe the multifaceted pain that service members feel after perpetrating, witnessing, or failing to prevent acts that conflict with their moral codes. More recently the term has been used to describe healthcare providers' pain stemming from their experiences serving on the frontlines of the healthcare system when: a medical error causes serious harm to patients, systems continuously impede their abilities to provide proper care, or providers assess that they have acted in ways that conflict with their professional ethics or oaths to "do no harm." This article explores moral injury risk at the intersection of military service and healthcare by examining challenges that military behavioral healthcare providers face. Leveraging moral injury definitions previously applied to service members (personal or witnessed transgressions) and in two healthcare contexts ("second victim" to adverse client outcomes and system-driven moral distress), as well as literature on ethical challenges in military behavioral health, this paper uncovers situations that may amplify military behavioral health providers' risks for moral injury. It concludes by offering policy and practice recommendations germane to military medicine aimed at alleviating pressures military behavioral healthcare providers face and mitigating moral injuries' potential ripple effects on provider wellness, retention and care quality.

Predictive modelling of deliberate self-harm and suicide attempts in young people accessing primary care: a machine learning analysis of a longitudinal study

PURPOSE

Machine learning (ML) has shown promise in modelling future self-harm but is yet to be applied to key questions facing clinical services. In a cohort of young people accessing primary mental health care, this study aimed to establish (1) the performance of models predicting deliberate self-harm (DSH) compared to suicide attempt (SA), (2) the performance of models predicting new-onset or repeat behaviour, and (3) the relative importance of factors predicting these outcomes.

METHODS

802 young people aged 12-25 years attending primary mental health services had detailed social and clinical assessments at baseline and 509 completed 12-month follow-up. Four ML algorithms, as well as logistic regression, were applied to build four distinct models.

RESULTS

The mean performance of models predicting SA (AUC: 0.82) performed better than the models predicting DSH (AUC: 0.72), with mean positive predictive values (PPV) approximately twice that of the prevalence (SA prevalence 14%, PPV: 0.32, DSH prevalence 22%, PPV: 0.40). All ML models outperformed standard logistic regression. The most frequently selected variable in both models was a history of DSH via cutting.

CONCLUSION

History of DSH and clinical symptoms of common mental disorders, rather than social and demographic factors, were the most important variables in modelling future behaviour. The performance of models predicting outcomes in key sub-cohorts, those with new-onset or repetition of DSH or SA during follow-up, was poor. These findings may indicate that the performance of models of future DSH or SA may depend on knowledge of the individual's recent history of either behaviour.

Significantly elevated phosphatidylethanol levels in recent suicide attempters, but not in depressed controls and healthy volunteers

INTRODUCTION

Suicide is a complex transdiagnostic phenomenon. It is strongly associated with, but not exclusive to major depressive disorder (MDD). Hazardous alcohol drinking has also been linked to an increased risk of suicidal behaviours, however, it is often underreported. The study aimed to evaluate whether an objective measure of chronic alcohol use, phosphatidylethanol (PEth) could be useful as a biomarker in clinical practice.

METHOD

ology. The present case-control multi-centric study recruited 156 participants into three study groups: 52 patients treated for major depressive disorder (MDD), 51 individuals immediately following a suicide attempt (SA), and 53 volunteers. Sociodemographic data, medical history, and laboratory data, including PEth concentrations and C-reactive protein levels, were collected from study participants.

RESULTS

PEth concentrations were the highest in suicide attempters (232,54 ± 394,01 ng/ml), followed by patients with MDD (58,39 ± 135,82 ng/ml), and the control group (24,45 ± 70,83 ng/ml) (Kruskall Wallis χ2 = 12.23, df = 2, p = .002). In a multinomial logistic regression model with adjustments, PEth concentration was able to predict belonging to suicide attempters' group, but not to depression group (p = .01). Suicide attempters were also more likely to underreport their recent alcohol consumption.

LIMITATIONS

We did not analyze SA methods, psychiatric comorbidity and several other factors that might be associated with PEth levels, such as body mass index, race, and haemoglobin levels. Sample recruited in hospital settings may not be representative of the whole population. The results of this adult-only study cannot be generalized to adolescents.

CONCLUSIONS

PEth levels in recent suicide attempters significantly exceeded those of patients with MDD and controls. Suicide attempters also were more likely to underreport their alcohol consumption when questioned about their consuption. PEth might be an interesting biomarker to evaluate individuals at risk of SA.

Suicide Screening Questionnaire-Self-Rating (SSQ-SR): Development, reliability, and validity in a clinical sample of Korean adults

The goal of the present study was to evaluate the psychometric properties of the Suicide Screening Questionnaire-Self-Rating (SSQ-SR). A 25-item SSQ-SR is a newly developed suicide screening tool that measures suicide risk factors, including a history of suicidal thoughts and behaviors (STBs), life stress, and mental health problems. To investigate the reliability and validity of the SSQ-SR, we conducted a longitudinal case-control study with adults with and without STBs in the past six months. A total of 176 participants were recruited through 12 hospital-based Crisis Response Centers across South Korea. At the baseline, we administered the SSQ-SR, the Beck Scale for Suicide Ideation (BSSI), and the Patient Health Questionnaire-9 (PHQ-9). In a 6-months follow-up, we investigated whether the participants engaged in suicidal ideation, plan, or attempt since the baseline assessment. As a result, the SSQ-SR demonstrated a strong internal consistency (Cronbach's alpha coefficient = 0.96). In addition, the total score of SSQ-SR had concurrent validity compared to the total scores of the BSSI and the PHQ-9. In comparing the suicidal groups with the control group, the ROC analysis indicated the optimal cut point at 31 with a sensitivity rate of 0.97 and a specificity rate of 0.98. Through explanatory factor analysis, two factors were identified: Mental Health and Environmental Factors and Active Suicidal Thoughts and Behaviors. The SSQ-SR total and sub-factor scores were prospectively associated with subsequent suicidal ideation, plan, and attempt. These findings support that the SSQ-SR is a promising tool in prospectively screening those who are at risk of suicidal thoughts, plans, and nonfatal attempts.

The association between vitamin D serum levels, supplementation, and suicide attempts and intentional self-harm

OBJECTIVES

The purpose of this study is to determine the associations between Vitamin D supplementation, 25(OH) blood serum levels, suicide attempts, and intentional self-harm in a population of veterans in the Department of Veterans Affairs (VA).

METHODS

A retrospective cohort study of US Veterans supplemented with Vitamin D. Veterans with any Vitamin D3 (cholecalciferol) or Vitamin D2 (ergocalciferol) fill between 2010 and 2018 were matched 1:1 to untreated control veterans having similar demographics and medical histories. Cox proportional hazards regression was used to estimate the time from the first Vitamin D3 (cholecalciferol) or Vitamin D2 (ergocalciferol) prescription fill to the first suicide attempt or intentional self-harm. Analyses were repeated in stratified samples to measure associations by race (Black or White), gender (male or female), blood levels (0-19 ng/ml, 20-39 ng/ml, and 40+ ng/ml), and average daily dosage.

RESULTS

Vitamin D3 and D2 supplementation were associated with a 45% and 48% lower risk of suicide attempt and self-harm ((D2 Hazard Ratio (HR) = 0.512, [95% CI, 0.457, 0.574]; D3 HR = 0.552, [95% CI, 0.511, 0.597])). Supplemented black veterans and veterans with 0-19 ng/ml vitamin D serum levels were at ~64% lower risk relative to controls (Black Veteran HR: 0.362 [95% CI: 0.298,0.440]; 0-19 ng/ml HR: 0.359 [95% CI: 0.215,0.598]). Supplementation with higher vitamin D dosages was associated with greater risk reductions than lower dosages (Log Average Dosage HR: 0.837 [95% CI: 0.779,0.900]).

CONCLUSIONS

Vitamin D supplementation was associated with a reduced risk of suicide attempt and self-harm in Veterans, especially in veterans with low blood serum levels and Black veterans.

An explainable predictive model for suicide attempt risk using an ensemble learning and Shapley Additive Explanations (SHAP) approach

Machine learning approaches have been used to develop suicide attempt predictive models recently and have been shown to have a good performance. However, those proposed models have difficulty interpreting and understanding why an individual has suicidal attempts. To overcome this issue, the identification of features such as risk factors in predicting suicide attempts is important for clinicians to make decisions. Therefore, the aim of this study is to propose an explainable predictive model to predict and analyse the importance of features for suicide attempts. This model can also provide explanations to improve the clinical understanding of suicide attempts. Two complex ensemble learning models, namely Random Forest and Gradient Boosting with an explanatory model (SHapley Additive exPlanations (SHAP)) have been constructed. The models are used for predictive interpretation and understanding of the importance of the features. The experiment shows that both models with SHAP are able to interpret and understand the nature of an individual's predictions with suicide attempts. However, compared with Random Forest, the results show that Gradient Boosting with SHAP achieves higher accuracy and the analyses found that history of suicide attempts, suicidal ideation, and ethnicity as the main predictors for suicide attempts.

A recent suicide attempt and the heartbeat: Electrophysiological findings from a trans-diagnostic cohort of patients and healthy controls

Suicidal behavior is influenced by a multitude of factors, making prediction and prevention of suicide attempts (SA) a challenge. A useful tool to uncover underlying pathophysiology or propose new therapy approaches are biomarkers, especially within the context of point-of-care tests. Heart rate variability (HRV) is a well-established biomarker of mental health, and measures the activity of the sympathetic and parasympathetic nervous system (PNS). Previous studies reported a correlation between lower PNS activity and suicidality. However, most studies involved participants from a healthy population, patients without history of suicide attempts, or patients with a single diagnosis. 52 in-patients with a recent suicide attempt (<6 months), and 43 controls without history of SA or psychiatric diagnoses confirmed study participation. The included patients age ranged between 18 and 65 years, 65% had psychiatric comorbidities. Patients with dementia, cognitive impairments, acute psychosis, chronic non suicidal self-harming behavior, or current electroconvulsive therapy were excluded. A 15-min resting state electrocardiography was recorded with two bipolar electrodes attached to the right and left insides of the wrists. The multiple regression analyses showed lower parasympathetic, and higher sympathetic activity in patients compared to controls. Partial correlation found a positive trend result between self-reported suicidality and the very low frequency band. ROC curve analysis revealed an acceptable to excellent clinical accuracy of HRV parameters. Therefore, HRV parameters could be reliable discriminative biomarkers between in-patients with a recent SA and healthy controls. One limitation is the lack of a control group consisting of in-patients without life-time suicidal ideation or attempts.

What happened to the predicted COVID-19-induced suicide epidemic, and why?

Two years ago, in the early stages of the COVID-19 pandemic, there were widespread and grim predictions of an ensuing suicide epidemic. Not only has this not happened but also by the end of 2021 in the majority of countries and regions with available data, the suicide rates had, if anything, declined. We discuss four reasons why the predictions of suicide models were exaggerated: (1) government intervention reduced the economic and mental costs of lockdowns, (2) the pandemic itself and lockdowns had less of an effect on mental health than assumed, (3) the evidence for a link between economic downturns, distress and suicide is weaker and less consistent than the models assumed and (4) predicting suicide is generally hard. Predictive models have an important place, but their strong modelling assumptions need to acknowledge the inherent high degree of uncertainty which has been further augmented by behavioural responses of pandemic management.

Combined effects of nitric oxide synthase 3 genetic variant and childhood emotional abuse on earlier onset of suicidal behaviours

Marked heterogeneity in suicide attempters has been observed, with earlier onset being linked to stronger heritability, more childhood maltreatment. Nitric oxide signalling system might be implicated in this relationship through its role in the stress response/adaptation. This study examined how NOS genetic variants and childhood maltreatment were associated with age at first suicide attempt (SA). Adult patients with SA history (N = 414) filled in the Childhood Trauma Questionnaire, and six functionally relevant NOS2 and NOS3 polymorphisms were genotyped. Analyses included χ², Mann-Whitney U tests, Kendall's regression, multivariate linear and Cox survival regressions, and a moderation analysis. The NOS3 promotor 27-bp variable number tandem repeat (VNTR) bb homozygous state and childhood emotional abuse were independently associated with earlier age at first SA, which was robust after controlling for confounders [regression coefficient - 3.975, 95% CI -6.980 - (-0.970), p = 0.010, and - 1.088, 95% CI -2.172 - (-0.004), p = 0.049]. No interaction was observed. In the Cox proportional hazards model for age at first SA, the hazard ratio for patients with childhood emotional abuse and NOS3 27-bp VNTR bb was 0.533 (95% CI 0.394-0.720, p < 0.001) compared to patients without. Intermediate scores were observed with either only the risk genotype or only childhood emotional abuse. A graded relationship was also observed for repeated SA, family history of SA, and severe SA history. These results are preliminary due to a low statistical power and call for replication and further characterization of the role of nitric oxide system in the susceptibility to early-onset SB.

Portability of natural language processing methods to detect suicidality from clinical text in US and UK electronic health records

Background

In the global effort to prevent death by suicide, many academic medical institutions are implementing natural language processing (NLP) approaches to detect suicidality from unstructured clinical text in electronic health records (EHRs), with the hope of targeting timely, preventative interventions to individuals most at risk of suicide. Despite the international need, the development of these NLP approaches in EHRs has been largely local and not shared across healthcare systems.

Methods

In this study, we developed a process to share NLP approaches that were individually developed at King's College London (KCL), UK and Weill Cornell Medicine (WCM), US - two academic medical centers based in different countries with vastly different healthcare systems. We tested and compared the algorithms' performance on manually annotated clinical notes (KCL: n = 4,911 and WCM = 837).

Results

After a successful technical porting of the NLP approaches, our quantitative evaluation determined that independently developed NLP approaches can detect suicidality at another healthcare organization with a different EHR system, clinical documentation processes, and culture, yet do not achieve the same level of success as at the institution where the NLP algorithm was developed (KCL approach: F1-score 0.85 vs. 0.68, WCM approach: F1-score 0.87 vs. 0.72).

Limitations

Independent NLP algorithm development and patient cohort selection at the two institutions comprised direct comparability.

Conclusions

Shared use of these NLP approaches is a critical step forward towards improving data-driven algorithms for early suicide risk identification and timely prevention.